What are the current challenges and progress of Autonomous Driving?
In recent years, advancements in autonomous driving technology have generated significant excitement, emerging as a new frontier for OEMs with substantial investments. Despite high public expectations, the market has probably yet to reach what could truly be considered ‘autonomous driving’.
Autonomous driving technology is generally categorized into levels 1 through 5, with levels 4 and 5 specifically aiming for complete autonomy without human intervention. While many enterprises are continuing to develop for levels 4 and 5, most vehicles are currently evaluated to fall between levels 1 and 3. Even as level 4 and 5 autonomous driving technologies are developed, numerous safety flaws continue to be uncovered. Apple has also withdrawn from the Apple Car project, which has lowered public expectations for the autonomous driving market. Additionally, according to a study by S&P Global Mobility, achieving complete level 5 autonomy by 2025 is expected to be challenging.
Fortunately, it seems that the efforts made so far have not been entirely in vain. Significant progress has been made in levels 1 to 3 (ADAS), with companies now focusing on level 3 and aiming for effective integration of driver assistance and autonomy, with lowered expectations. Furthermore, numerous companies are diving into the robo-taxi industry, signaling significant advancement in autonomous driving. Here’s a snapshot of the leading players shaping this market.
Tesla
Tesla, a major player in the autonomous driving industry, has recently struggled to meet expectations. Last year, Tesla recalled 2 million US vehicles due to concerns over its ‘Autopilot’ feature. Despite this recall, there were 20 collisions related to this feature, prompting US transportation authorities to launch investigations. Additionally, Tesla received poor ratings in a series of tests conducted by the Insurance Institute for Highway Safety. In response, CEO Elon Musk announced the launch of the ‘Cybercab’ robotaxi and affordable models, aiming to once again raise public’s expectations for Tesla. Tesla has amassed vast amounts of data over the years and continues to lead the market steadily based on this data. It remains to be seen whether Tesla’s robotaxi can rejuvenate the stagnating autonomous driving market with its enormous data pool.
*Source: Tesla
Cruise
Another front runner in the industry, GM, also faced challenges last year with its robotaxi. GM’s subsidiary, Cruise, introduced robotaxis in several states including California last year, but a series of accidents caused by unstable technology led to the complete suspension of Cruise taxi operations. Consequently, Cruise recorded a total operating loss of $3.48 billion in 2023, prompting GM to reduce its investment in Cruise by about $1 billion. Yet, despite these setbacks, Cruise remains committed to autonomous driving. Cruise aims to return to US roads within 1-2 years, and has launched a manually driven vehicle program to create maps and collect road data. Training data is crucial for resuming their service. Elshenawy from Cruise highlighted the significance of data in the autonomous driving market, stating “Combining road and simulation data not only improves older models but also aids in training new AI models”. Can Cruise overcome this challenge and leap back to the forefront of the autonomous driving field?
*Source: Cruise
Waymo
Following the suspension of Cruise’s robotaxi service, another player in the field, Waymo, is consolidating its market dominance. The fact that Waymo has raised $1.05 billion for autonomous driving technology seems to validate this. As a subsidiary of Google’s Alphabet, Waymo successfully launched robotaxi services in San Francisco and expanded to Los Angeles in March of this year. The announcement garnered over 50,000 sign-ups on Waymo’s waiting list, reflecting high anticipation. Yet, concerns and anger have also arisen, exemplified by a recent incident in San Francisco’s Chinatown where a Waymo vehicle caught fire due to human action. Also, there were collision accidents by its vehicle, underscoring imperfections in autonomous driving. Despite these setbacks, Waymo continues to operate its robotaxi, and the key to Waymo’s sustained operation likely lies in its driving data. Waymo’s autonomous driving system, ‘Waymo Driver’, has accumulated over 40 million miles (over 64 million km) of driving data to date, equivalent to driving round trip to the moon 80 times. It will be intriguing to observe whether Waymo can sustain its dominant position in the future.
*Source: Waymo
What is the most important factor for technological growth in the autonomous driving market?
The autonomous driving market continues to evolve with repeated failures. Among them, there is lively discussion about sensors, one of the most crucial elements acting as the “eyes” of autonomous driving. There is still no definitive methodology between computer vision and LiDAR, and everyone is exploring them to devise the safest technology for autonomous driving. However, whether it’s computer vision or LiDAR, the undeniable fact is that data is the foundation for implementing them.
The efficacy of autonomous systems heavily relies on the quality and quantity of data they absorb through AI deep learning. Initially, companies believed in ‘data maximization,’ assuming that more data would lead to smarter systems. However, Cruise found that only 1% of data collected during San Francisco road tests was useful. Due to the volume and complexity of data, enterprises are now prioritizing selectively retaining and organizing valuable data. The focus is shifting towards gathering high-quality data and efficiently labeling them for superior performance.
Therefore, any enterprise aiming to make a significant impact in the autonomous driving market will need well-refined data. CrowdWorks AI, known for its years of expertise and a proven track record in AD training data, could be a valuable resource in advancing your autonomous driving initiatives.
For more details or to explore how we can assist, please visit the link below:
